Automated Generation of Individual and Population Clinical Pathways with the OMOP Common Data Model

Clinical pathways represents the sequence of interventions from which the patients benefit during their encounters with health care structures. There are several complex issues which make it difficult to represent these pathways (e.g. high numbers of patients, heterogeneity of variables).

We developed a tool to automate the representation of clinical pathways, from an individual and population points of view, and based on the OMOP CDM. The tool implemented the Sankey diagram in three stages: (i) data extraction, (ii) generation of individual sequence of steps and (iii) aggregation of sequence to obtain the population-level diagram. We tested the tool with three surgery procedures: the total hip replacement, the coronary bypass and the transcatheter aortic valve implantation.

The tool provided different ways of visualizing pathways depending on the question asked: a pathway before a surgery, the pathway of deceased patients or the complete pathway with different steps of interest.

We proposed a tool automating the representation of the clinical pathways, and reducing complexity of visualization. Representations are detailed from an individual and population points of view. It has been tested with three surgical procedures. The tool functionalities will be extended to cover a greater number of use cases.